SummaryAcute myeloid leukaemia (AML) has been proposed to arise from the collaboration between two classes of mutation, a class I, or proliferative, mutation and a class II, or blocking, mutation. A limitation of this so-called 'two-hit' hypothesis has been the lack of identifiable proliferative and blocking mutations in most AML cases. However, it is now known that the CBFb-MYH11 fusion gene in AML and inv(16), by disrupting the normal transcription factor activity of core binding factor (CBF), functions as a class II mutation. In addition, nearly 70% of patients with AML and inv(16) are known to possess mutually exclusive mutations of the receptor tyrosine kinases (RTKs), c-KIT and FLT3, as well as RAS genes, that provide a class I, or proliferative, signal. AML and inv(16), therefore, is one of the best understood of the acute leukaemias at the genetic level and so provides a paradigm for the 'two-hit' hypothesis of leukaemogenesis. This paper reviews the recent advances in the molecular pathology of AML and inv(16) and discusses possible therapeutic implications of the current pathogenetic model. Keywords: acute myeloid leukaemia, c-KIT, core binding factor, core binding factorb-MYH11, pathogenesis.Acute myeloid leukaemia (AML) is a heterogeneous clonal disorder with individual cases exhibiting variability in clinical presentation, cellular morphology, therapeutic response and overall prognosis. However, although this heterogeneity also extends to the underlying mutations, the end effect is similar, in that each patient's genotype confers deregulated proliferation, impaired differentiation and a survival advantage for the leukaemic cells. The number of known mutations associated with AML continues to grow at an unprecedented pace, with over 300 different chromosomal translocations and other mutational events having been described. It is obvious, therefore, that there are many more leukaemic genotypes than phenotypes. In an attempt to provide a unified molecular theme to explain how different mutations can generate essentially similar phenotypes, Gilliland (2001) has proposed a 'two-hit' model for leukaemogenesis. The basis of the hypothesis is that AML is the consequence of a collaboration between at least two broad classes of mutation; class I mutations that confer a proliferative and/or survival advantage to cells (e.g. BCR-ABL and oncogenic RAS) and class II mutations that primarily impair haematopoietic differentiation and subsequent cellular apoptosis (e.g. CBFb-MYH11 and PML-RARa fusion genes) ( Table I). At its simplest, the model predicts that AML results from the combined effects of only two mutations, one from each class. However, a limitation of the model has been the lack of identifiable class I and class II mutations in the majority of AML cases.Recently, however, it has become apparent that mutations of receptor tyrosine kinases (RTKs) class III and RAS frequently provide the 'missing' proliferative signal in AML. The purpose of this review is to discuss these findings in the context of the pat...